← Back to team overview

kicad-developers team mailing list archive

Re: eeschema fields moved in version 3230


On 01/06/2012 01:21 PM, jean-pierre charras wrote:
> Le 06/01/2012 16:14, Dick Hollenbeck a écrit :
>> On 01/05/2012 05:05 PM, Dick Hollenbeck wrote:
>>> On 01/05/2012 02:22 PM, jean-pierre charras wrote:
>>>> Le 05/01/2012 20:52, Dick Hollenbeck a écrit :
>>>>> Jean-Pierre,
>>>>> All my EESCHEMA part "texts fields" were moved when using revision 3230 and later, on
>>>>> screen.  These text fields are now shifted up in most cases.  I can send jpegs if you are
>>>>> not familiar with the problem.
>>>>> On one schematic, which consists of 17 sheets, it seems like a lot of work to "fix" what
>>>>> was already a good schematic.  I don't know if the changes in
>>>>>       eeschema/sch_field.cpp, rev 3230
>>>>> were worth this inconvenience.  I wonder if it was your intention to actually move all
>>>>> fields of all schematics ever created with KiCad before 3230?
>>>>> Thanks,
>>>>> Dick
> I just committed a change in sch_field.cpp,
> to go back to old behavior for field texts justifications (I am thinking this patch works in all cases).
> Note: only vertical texts have a modified behavior.
> This option is not currently activated.
> To test it, change the line
>   to
> in sch_field.cpp
> Set USE_TEXT_JUSTIFY_INITIAL_BEHAVIOR to 0 to use  a justification relative to the text itself
> (the current behavior)
> The justification relative to the horizontal text.
> Se to 1 to go back to the initial Eeschema behavior.
> The initial behavior is:
>    For vertical texts, exchange the horizontal and the vertical justification
>    The idea is to keep the justification always left or top for instance, no matter the text orientation
> Unfortunately, this is very tricky when you want to change a text field justification,
> when the field and the component are both rotated 90.0 degrees
> This is the reason I changed the justification of rotated texts.
> For instance, using the old behavior:
> * load a component like "R" from the device lib (the value is rotated 90.0 degrees).
> * Rotate the component 90.0 degrees.
> * change "R" to a bigger text.
> and try to change a justification...
> (Try the same thing with the new behavior).
> However the new behavior breaks schematics that use many components having fields rotated,
> and using a not centered justification.
> So we make a decision: old or new behavior?
> I need your opinion.

My original question was whether the value in the change was worth the cost.   Now that I
know more about the original problem, I will be better positioned to offer an opinion on
it myself.  Just knowing whether it is was a new bug or a fix, is a good start on
gathering all the data.  It sounds like it truly was a fix, but one that broke the
appearance of older schematics.

Give me some time to formulate an opinion, and thanks for the opportunity to offer one. 
There may be additional options I can offer also.

Once again, thanks for your valuable time, Jean Pierre.